US11740603B2 - Power load prediction method and apparatus, and storage medium - Google Patents

Power load prediction method and apparatus, and storage medium Download PDF

Info

Publication number
US11740603B2
US11740603B2 US17/614,401 US201917614401A US11740603B2 US 11740603 B2 US11740603 B2 US 11740603B2 US 201917614401 A US201917614401 A US 201917614401A US 11740603 B2 US11740603 B2 US 11740603B2
Authority
US
United States
Prior art keywords
power load
data
time scale
time
load
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
US17/614,401
Other languages
English (en)
Other versions
US20220214655A1 (en
Inventor
Wen Tao HUA
Jing Li
Hao Liu
Dan Wang
Ang Li
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Siemens AG
Original Assignee
Siemens AG
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Siemens AG filed Critical Siemens AG
Assigned to SIEMENS LTD., CHINA reassignment SIEMENS LTD., CHINA ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: LI, ANG, HUA, Wen Tao, LI, JING, LIU, HAO, WANG, DAN
Assigned to SIEMENS AKTIENGESELLSCHAFT reassignment SIEMENS AKTIENGESELLSCHAFT ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: SIEMENS LTD., CHINA
Publication of US20220214655A1 publication Critical patent/US20220214655A1/en
Application granted granted Critical
Publication of US11740603B2 publication Critical patent/US11740603B2/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
    • G05B19/042—Programme control other than numerical control, i.e. in sequence controllers or logic controllers using digital processors
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/20—Pc systems
    • G05B2219/26—Pc applications
    • G05B2219/2639—Energy management, use maximum of cheap power, keep peak load low
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Definitions

  • Embodiments of the present invention generally relate to the energy field, and in particular to a power load prediction method, apparatus, cloud platform, server and storage medium.
  • Power industry is the main infrastructure in the energy sector, and plays an important role in the development of industry and the quality of life.
  • Power load for example, the power of equipment such as transformers, etc.
  • Continuous overload will cause damage to electrical equipment, for example, transformers.
  • it is necessary to monitor the power load in advance.
  • power load is usually predicted based on the growth rate, and the growth rate is calculated based on user tags.
  • the tags of registered users at the administration for power supply are relatively fixed, and these tags cannot reflect the latest conditions of users. Therefore, power load prediction based on the growth rate severely limits the prediction accuracy.
  • a power load prediction method is proposed in one aspect of the embodiments of the present invention, and a power load prediction apparatus, a cloud platform, a server, a storage medium and a computer program product are provided in another aspect, which are used to improve accuracy of power load prediction.
  • a power load prediction method proposed in the embodiments of the present invention comprises: acquiring historical power load data in a one-dimensional time series for a set time length, which consists of data corresponding to each time point; converting the historical power load data in the one-dimensional time series to a three-dimensional matrix consisting of set time scales, days in each time scale, and time points in each day; dividing the historical power load data of the three-dimensional matrix into at least one operation mode depending on the size of the time scales; and, in each operation mode, taking the time scale as the unit, deriving the value band of power load data of each day in the operation mode in the next time scale to be predicted based on the historical power load data in each time scale.
  • Another power load prediction apparatus proposed in the embodiments of the present invention comprises: at least one memory and at least one processor, wherein, the at least one memory is used to store a computer program; the at least one processor is used to call the computer program stored in the at least one memory, to execute the power load prediction method described in any of the above implementations.
  • a cloud platform or server proposed in the embodiments of the present invention comprises a power load prediction apparatus described in any of the above implementations.
  • a computer-readable storage medium proposed in the embodiments of the present invention has a computer program stored thereon; it is characterized in that the computer program can be executed by a processor and implement the power load prediction method described in any of the above implementations.
  • a computer program product proposed in the embodiments of the present invention is stored on a computer-readable storage medium, comprises a computer program instruction, which, when executed, causes at least one processor to execute the power load prediction method described in any of the above implementations.
  • FIG. 1 is an example flowchart of a power load prediction method in the embodiments of the present invention.
  • FIG. 2 is schematic diagram of the method for deriving the power load data in the next time scale to be predicted based on the historical power load data in each time scale.
  • FIG. 3 is a schematic diagram of the value band of power load data of each day in an operation mode in the next time scale to be predicted in one example of the present invention.
  • FIG. 4 is an example structural diagram of a power load prediction apparatus in the embodiments of the present invention.
  • FIG. 5 is a schematic structural diagram of the data prediction module shown in FIG. 3 in the embodiments of the present invention.
  • FIG. 6 is an example flowchart of another power load prediction apparatus in the embodiments of the present invention.
  • a power load prediction method proposed in the embodiments of the present invention comprises: acquiring historical power load data in a one-dimensional time series for a set time length, which consists of data corresponding to each time point; converting the historical power load data in the one-dimensional time series to a three-dimensional matrix consisting of set time scales, days in each time scale, and time points in each day; dividing the historical power load data of the three-dimensional matrix into at least one operation mode depending on the size of the time scales; and, in each operation mode, taking the time scale as the unit, deriving the value band of power load data of each day in the operation mode in the next time scale to be predicted based on the historical power load data in each time scale.
  • dividing the historical power load data of the three-dimensional matrix into at least one operation mode depending on the size of the time scales comprises: when the time scale is smaller than a set first threshold, dividing the historical power load data of the three-dimensional matrix into one operation mode; when the time scale is greater than or equal to the first threshold, clustering the historical power load data of the three-dimensional matrix in the unit of day, to obtain a plurality of clusters, each corresponding to one operation mode.
  • the method further comprises: calculating the ratio of the days in each cluster to the total number of days, abandoning clusters with the ratio smaller than a set second threshold, and at the same time abandoning the operation modes corresponding to the clusters.
  • time scales smaller than the first threshold include: month or week; time scales greater than the first threshold include: year.
  • the method further comprises: obtaining external data in the time period corresponding to the load change curve, wherein the external data comprises weather data and/or production plan data; calculating the relevance of the external data to the load change curve; determining that the external data is relevant to the load change curve when the relevance is greater than a set third threshold; determining that the external data is relevant to power load data when the proportion of all load change curves with relevant external data in the corresponding time periods reaches a set fourth threshold; and deriving the dominant load curve of the next time scale to be predicted based on all the dominant load curves and all the load change curves, specifically: deriving the dominant load curve of the next time scale to be predicted based on all the dominant load curves, all the load change curves and the current external data.
  • the data prediction module comprises: a first unit, used to, in each operation mode, taking the time scale as the unit, determine the representative load value of the historical power load data at each time point in a day in each time scale, to obtain a dominant load curve; a second unit, used to, in each operation mode, for every two adjacent time scales, calculate the change in the dominant load curve of the latter time scale compared with the preceding time scale, to obtain a load change curve; a third unit, used to, in each operation mode, derive the dominant load curve of the next time scale to be predicted based on all the dominant load curves and all the load change curves; a fourth unit, used to, in each operation mode, determine a confidence interval of the dominant load curve of the next time scale to be predicted based on the values of the historical power load data in each time scale; and a fifth unit, used to, in each operation mode, obtain the value band of power load data of each day in the operation mode in the next time scale to be predicted based on the dominant load curve and the confidence interval of the
  • the operation mode dividing module divides the historical power load data of the three-dimensional matrix into one operation mode when the time scale is smaller than a set first threshold; and clusters the historical power load data of the three-dimensional matrix in the unit of day, to obtain a plurality of clusters, each corresponding to one operation mode, when the time scale is greater than or equal to the first threshold.
  • the apparatus further comprises: a simplifying module, used to calculate the ratio of the days in each cluster to the total number of days, abandon clusters with the ratio smaller than a set second threshold, and at the same time abandon the operation modes corresponding to the clusters.
  • a simplifying module used to calculate the ratio of the days in each cluster to the total number of days, abandon clusters with the ratio smaller than a set second threshold, and at the same time abandon the operation modes corresponding to the clusters.
  • time scales smaller than the first threshold include: month or week; time scales greater than the first threshold include: year.
  • the time scale is one month or one week; the data conversion module firstly converts the historical power load data in the one-dimensional time series to an initial three-dimensional matrix consisting of years, days in each year and time points in each day, and then converts the initial three-dimensional matrix to a three-dimensional matrix consisting of the time scales, days in each time scale and time points in each day.
  • the data prediction module further comprises: a sixth unit, used to obtain external data in the time period corresponding to each load change curve obtained by the second unit, wherein the external data comprises weather data and/or production plan data; calculate the relevance of the external data to the load change curve; and determine that the external data is relevant to the load change curve when the relevance is greater than a set third threshold; and a seventh unit, used to determine that the external data is relevant to power load data when the proportion of all load change curves with relevant external data in the corresponding time periods reaches a set fourth threshold; and the fifth unit derives the dominant load curve of the next time scale to be predicted based on all the dominant load curves, all the load change curves and the current external data.
  • Another power load prediction apparatus proposed in the embodiments of the present invention comprises: at least one memory and at least one processor, wherein, the at least one memory is used to store a computer program; the at least one processor is used to call the computer program stored in the at least one memory, to execute the power load prediction method described in any of the above implementations.
  • a cloud platform or server proposed in the embodiments of the present invention comprises a power load prediction apparatus described in any of the above implementations.
  • a computer-readable storage medium proposed in the embodiments of the present invention has a computer program stored thereon; it is characterized in that the computer program can be executed by a processor and implement the power load prediction method described in any of the above implementations.
  • a computer program product proposed in the embodiments of the present invention is stored on a computer-readable storage medium, comprises a computer program instruction, which, when executed, causes at least one processor to execute the power load prediction method described in any of the above implementations.
  • data in a one-dimensional time series is converted to data of a three-dimensional matrix based on a time scale, then the data is divided into operation modes based on the data of the three-dimensional matrix, and, in each operation mode, power load data of the next time scale is predicted based on historical power load data in each time scale.
  • Accuracy of power load prediction can be improved because the prediction method is based on historical power load data rather than user tags, and uses a technique based on data of a three-dimensional matrix for prediction instead of techniques based on one-dimensional time series for time series analysis that tend to lose data characteristics.
  • the time scale is taken as the unit to determine the dominant load curve of the historical load data in each time scale, and for every two adjacent time scales, the change curve of the dominant load curve of the latter time scale compared with the preceding time scale is calculated, thus deriving the dominant load curve of the next time scale to be predicted; the confidence interval of the dominant load curve of the next time scale to be predicted is determined based on the values of the historical power load data in each time scale; and the value band of power load data of each day in the operation mode in the next time scale to be predicted is obtained based on the dominant load curve and the confidence interval of the next time scale to be predicted.
  • Statistical techniques are mainly used in the process, and therefore it is easy to implement and can ensure prediction accuracy.
  • data in the converted three-dimensional matrix is clustered in the unit of day for long time scales, thereby making it possible to classify days with similar characteristics as one cluster, corresponding to one operation mode. Then, in each operation mode, the historical power load data in each time scale is used to predict the power load in the next time scale, which further improves prediction accuracy.
  • the ratio of the days in each cluster to the total number of days is calculated, and clusters with the ratio smaller than a set second threshold and at the same time the operation modes corresponding to the clusters are abandoned, which reduces the calculation work for useless operation modes and saves resources for calculation and storage.
  • the historical power load data in the one-dimensional time series is converted to a three-dimensional matrix based on a longer time scale, thus providing higher applicability of the three-dimensional matrix, i.e., the complex conversion from one-dimensional data to a three-dimensional data needs to be performed only once before data in both longer and shorter time scales can be easily further converted through the three-dimensional matrix based on a long time scale, without the need to perform again the complex conversion from one-dimensional data to three-dimensional data, thereby reducing the difficulty in data processing.
  • power load data for example, power data
  • time series i.e., power load data
  • power load data are data in one-dimensional time series.
  • some characteristics of historical data may be lost because time series analysis is based only on one-dimensional data for prediction.
  • through creative work it is conceived to convert data in one-dimensional time series to data of a three-dimensional matrix, and then to predict power load based on data of the three-dimensional matrix.
  • power load may be predicted based on different time scales (for example, year, month, week, etc.).
  • the power load data in the coming year is predicted based on the power load data in previous years (for example, two years); for another example, the power load data in the coming month is predicted based on the power load data in the previous months (for example, 24 months); for yet another example, the power load data in the coming week is predicted based on the power load data in previous weeks (for example, 52 weeks).
  • FIG. 1 is an example flowchart of a power load prediction method 1 in the embodiments of the present invention. As shown in FIG. 1 , the method 1 may comprise the following steps:
  • Step S 102 acquiring historical power load data in a one-dimensional time series for a set time length, which consists of data corresponding to each time point.
  • the set time length may be determined based on actual needs.
  • the set time length may be 2 years when the time scale is one year.
  • Step S 104 converting the historical power load data in the one-dimensional time series to a three-dimensional matrix consisting of set time scales, days in each time scale, and time points in each day.
  • the historical power load data in the one-dimensional time series may be converted to a three-dimensional matrix consisting of years, days in each year, and time points in each day in this step. For example, if data are collected at 96 time points in a day, there are 365 days in a year, and the time length of the historical power load data is 2 years, a matrix can be obtained with 2, 365 and 96 points respectively on the three dimensions.
  • the historical power load data in the one-dimensional time series may be converted to a three-dimensional matrix consisting of months, days in each month, and time points in each day in this step; if the time scale is one week, the historical power load data in the one-dimensional time series can be converted to a three-dimensional matrix consisting of weeks, days in each week, and time points in each day.
  • the historical power load data in the one-dimensional time series may be firstly converted to an initial three-dimensional matrix consisting of years, days in each year and time points in each day, and then the initial three-dimensional matrix is converted to a three-dimensional matrix consisting of the time scales, for example, months (or weeks), days in each time scale, for example, month (or week), and time points in each day.
  • Step S 106 dividing the historical power load data of the three-dimensional matrix into at least one operation mode depending on the size of the time scale.
  • an operation mode may be understood as a cluster of data with similar characteristics. For example, if power load data curves are plotted taking day as the unit, the power load data of all the days with a similar curve shape is in one cluster, i.e., one operation mode.
  • the historical power load data of the three-dimensional matrix may be clustered taking day as the unit. In this way, all the days with similar characteristics can be divided into one cluster, thereby obtaining a plurality of different clusters, each corresponding to one operation mode.
  • the time scale is short, being for example, one month or one week, only one operation mode may be defined, and the data in a whole month or a whole week is used to predict the data in the subsequent entire month or entire week.
  • the historical power load data of the three-dimensional matrix may be divided into one operation mode when the time scale is smaller than a set first threshold; and the historical power load data of the three-dimensional matrix is clustered in the unit of day, to obtain a plurality of clusters, each corresponding to one operation mode, when the time scale is greater than or equal to the first threshold.
  • the first threshold may be set based on actual conditions, for example, to a reasonable value such as 50 days, 60 days, 100 days, 200 days, etc., which is not limited here.
  • the ratio of the days in each cluster to the total number of days can be further calculated, and clusters with the ratio smaller than a set second threshold and at the same time the operation modes corresponding to the clusters are abandoned. For example, for two years with a total of 730 days, if a cluster comprises a few days, such as only one day, or three or five days, or even ten days or half a month, the cluster and the corresponding operation mode can be abandoned.
  • the second threshold may be set based on actual conditions, which is not limited here.
  • Step S 108 in each operation mode, taking the time scale as the unit, deriving the value band of the power load data of each day in the operation mode in the next time scale to be predicted based on the historical power load data in each time scale.
  • a trained neural network prediction model may be used for prediction, i.e., the historical power load data in each time scale is used as the input to the neural network prediction model, and the value band of the power load date of each day in the operation mode in the next time scale to be predicted output from the neural network prediction model is received.
  • each step in the method 2 as shown in FIG. 2 may be performed for each operation mode.
  • Step S 202 taking the time scale as the unit, determining the representative load value of the historical power load data at each time point in a day in each time scale, to obtain a dominant load curve.
  • the value at the 1st time point of each day in a month may be used to determine the representative load value at the 1st time point
  • the value at the 2nd time point of each day in a month may be used to determine the representative load value at the 2nd time point
  • the value at the 3rd time point of each day in a month may be used to determine the representative load value at the 3rd time point
  • the value at the 96th time point of each day in a month is used to determine the representative load value at the 96th time point, thereby obtaining a dominant load curve consisting of 96 representative load values.
  • Step S 204 for every two adjacent time scales, calculating the change in the dominant load curve of the latter time scale compared with the preceding time scale, to obtain a load change curve.
  • Step S 206 deriving the dominant load curve of the next time scale to be predicted based on all the dominant load curves and all the load change curves.
  • Step S 208 determining a confidence interval of the dominant load curve of the next time scale to be predicted based on the values of the historical power load data in each time scale.
  • Step 210 obtaining the value band of power load data of each day in the operation mode in the next time scale to be predicted based on the dominant load curve and the confidence interval of the next time scale to be predicted.
  • FIG. 3 is a schematic diagram of the value band of power load data of each day in an operation mode in the next time scale to be predicted in one example.
  • a day with 96 time points is taken as an example, and it can be seen that the curve with dots is the dominant load curve in the next time scale to be predicted, while the two curves above and below it are respectively the upper and the lower confidence intervals.
  • the method may further comprise: obtaining external data in the time period corresponding to the load change curve, wherein the external data comprises weather data and/or production plan data; calculating the relevance of the external data to the load change curve; determining that the external data is relevant to the load change curve when the relevance is greater than a set third threshold; and determining that the external data is relevant to power load data when the proportion of all load change curves with relevant external data in the corresponding time periods reaches a set fourth threshold.
  • the dominant load curve of the next time scale to be predicted may be derived based on all the dominant load curves, all the load change curves and the current external data.
  • the third threshold and the fourth threshold may be set based on actual conditions, which are not limited here.
  • the power load prediction methods in the embodiments of the present invention are described in detail above.
  • the power load prediction apparatuses in the embodiments of the present invention will then be described below.
  • the apparatuses in the embodiments of the present invention may be used to implement the above methods in the embodiments of the present invention. Please refer to the corresponding description of the method embodiments above for item not disclosed in detail for the apparatus embodiments of the present invention, which will not be repeated here.
  • FIG. 4 is an example structural diagram of a power load prediction apparatus in the embodiments of the present invention.
  • the apparatus 4 may comprise: a one-dimensional data acquiring module 410 , a data conversion module 420 , an operation mode dividing module 430 , and a data prediction module 440 .
  • the one-dimensional data acquiring module 410 is used to acquire historical power load data in a one-dimensional time series for a set time length, which consists of data corresponding to each time point.
  • the data conversion module 420 is used to convert the historical power load data in the one-dimensional time series to a three-dimensional matrix consisting of set time scales, days in each time scale, and time points in each day. For example, if the time scale is greater than or equal to a set first threshold, being for example, one year, the data conversion module 420 can convert the historical power load data in the one-dimensional time series to a three-dimensional matrix consisting of years, days in each year, and time points in each day.
  • a set first threshold being for example, one year
  • the data conversion module 420 can convert the historical power load data in the one-dimensional time series to a three-dimensional matrix consisting of months, days in each month, and time points in each day; if the time scale is one week, the data conversion module 420 can convert the historical power load data in the one-dimensional time series to a three-dimensional matrix consisting of weeks, days in each week, and time points in each day.
  • the historical power load data in the one-dimensional time series may be firstly converted to an initial three-dimensional matrix consisting of years, days in each year and time points in each day, and then the initial three-dimensional matrix is converted to a three-dimensional matrix consisting of the time scales, for example, months (or weeks), days in each time scale, for example, month (or week), and time points in each day.
  • the operation mode dividing module 430 is used to divide the historical power load data of the three-dimensional matrix into at least one operation mode depending on the size of the time scale.
  • the operation mode dividing module 430 can divide the historical power load data of the three-dimensional matrix into one operation mode when the time scale is smaller than a set first threshold; and cluster the historical power load data of the three-dimensional matrix in the unit of day, to obtain a plurality of clusters, each corresponding to one operation mode, when the time scale is greater than or equal to the first threshold.
  • the apparatus 3 may further comprise a simplifying module (not shown in FIG. 4 ), used to calculate the ratio of the days in each cluster to the total number of days, abandon clusters with the ratio smaller than a set second threshold, and at the same time abandon the operation modes corresponding to the clusters.
  • the data prediction module 440 is used to, in each operation mode, taking the time scale as the unit, derive the value band of the power load data of each day in the operation mode in the next time scale to be predicted based on the historical power load data in each time scale.
  • the data prediction module 440 may be implemented in several forms.
  • the data prediction module 440 may have the structure shown by the solid lines in FIG. 5 .
  • the data prediction module 440 may comprise: a first unit 441 , a second unit 442 , a third unit 443 , a fourth unit 444 , and a fifth unit 445 .
  • the first unit 441 is used to, in each operation mode, taking the time scale as the unit, determine the representative load value of the historical power load data at each time point in a day in each time scale, to obtain a dominant load curve.
  • the second unit 442 is used to, in each operation mode, for every two adjacent time scales, calculate the change in the dominant load curve of the latter time scale compared with the preceding time scale, to obtain a load change curve.
  • the third unit 443 is used to, in each operation mode, derive the dominant load curve of the next time scale to be predicted based on all the dominant load curves and all the load change curves.
  • the fourth unit 444 is used to, in each operation mode, determine a confidence interval of the dominant load curve of the next time scale to be predicted based on the values of the historical power load data in each time scale.
  • the fifth unit 445 is used to, in each operation mode, obtain the value band of power load data of each day in the operation mode in the next time scale to be predicted based on the dominant load curve and the confidence interval of the next time scale to be predicted.
  • the data prediction module 440 as shown in FIG. 5 may further comprise a sixth unit 446 and a seventh unit 447 , as shown by the dotted lines in FIG. 4 .
  • the sixth unit 446 is used to obtain external data in the time period corresponding to each of the load change curves obtained by the second unit 442 , wherein the external data comprises weather data and/or production plan data; calculate the relevance of the external data to the load change curve; and determine that the external data is relevant to the load change curve when the relevance is greater than a set third threshold.
  • the seventh unit 447 is used to determine that the external data is relevant to power load data when the proportion of all load change curves with relevant external data in the corresponding time periods reaches a set fourth threshold.
  • the fifth unit 445 can derive the dominant load curve of the next time scale to be predicted based on all the dominant load curves, all the load change curves and the current external data.
  • FIG. 6 is an example flowchart of another power load prediction apparatus in the embodiments of the present invention.
  • the apparatus may comprise: at least one memory 61 and at least one processor 62 .
  • some other components for example, communication ports, etc., may also be comprised. These components communicate via a bus.
  • the at least one memory 61 is used to store a computer program.
  • the computer program may be understood as comprising each of the modules of the power load prediction apparatus as shown in FIG. 4 .
  • the at least one memory 61 may also store an operating system, etc.
  • the operating system may be but is not limited to: an Android operating system, a Symbian operating system, a Windows operating system, a Linux operating system, etc.
  • the at least one processor 62 is used to call the computer program stored in the at least one memory 61 to execute the power load prediction methods described in the embodiments of the present invention.
  • the processor 62 may be a CPU, a processing unit/module, and ASIC, a logic module, a programmable gate array, etc. It can receive and send data through the communication ports.
  • a server or a server cluster, or a cloud platform, etc. comprising the power load prediction apparatus shown in FIG. 2 or 3 above is also provided in the embodiments of the present invention.
  • a hardware module may comprise specially designed permanent circuits or logic devices (for example, dedicated processors, such as FPGA or ASIC) to complete specific operations.
  • a hardware module may also comprise programmable logic devices or circuits temporarily configured by software (for example, general-purpose processors or other programmable processors) for performing specific operations. Whether to specifically use mechanical methods or dedicated permanent circuits or temporarily configured circuits (such as software configuration) to implement hardware modules may be determined according to cost and schedule considerations.
  • the embodiments of the present invention also provide computer software that can be executed on a server or a server cluster or a cloud platform.
  • the computer software can be executed by a processor and implement the power load prediction methods described in the embodiments of the present invention.
  • a computer-readable storage medium is also provided in the embodiments of the present invention, which has a computer program stored thereon, which can be executed by a processor and implement the power load prediction methods described in the embodiments of the present invention.
  • a system or device equipped with a storage medium may be provided, and the software program code for implementing the functions of any of the above implementations is stored on the storage medium, so that a computer (or CPU or MPU) of the system or device reads and executes the program code stored in the storage medium.
  • the operating system operating on the computer may also be used to perform part or all of the actual operations through instructions based on the program code.
  • the program code read from the storage medium may be written to the memory provided in an expansion board inserted into the computer or to the memory provided in an expansion unit connected to the computer, and then the program code-based instructions cause the CPU, etc. mounted on the expansion board or the expansion unit to perform part and all of the actual operations, so as to implement the functions of any of the above embodiments.
  • Implementations of the storage media used to provide the program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards and ROMs.
  • the program code may be downloaded from a server computer via a communication network.
  • a computer program product is provided in the embodiments of the present invention, which is stored on a computer-readable storage medium, and comprises a computer program instruction, which, when executed, causes at least one processor to execute the power load prediction method described in any of the above implementations.
  • data in a one-dimensional time series is converted to data of a three-dimensional matrix based on a time scale, then the data is divided into operation modes based on the data in the three-dimensional matrix, and, in each operation mode, power load data of the next time scale is predicted based on historical power load data in each time scale.
  • Accuracy of power load prediction can be improved because the prediction method is based on historical power load data rather than user tags, and uses a technique based on data of a three-dimensional matrix for prediction instead of techniques based on one-dimensional time series for time series analysis that tend to lose data characteristics.
  • the time scale is taken as the unit to determine the dominant load curve of the historical load data in each time scale, and for every two adjacent time scales, the change curve of the dominant load curve of the latter time scale compared with the preceding time scale is calculated, thus deriving the dominant load curve of the next time scale to be predicted; the confidence interval of the dominant load curve of the next time scale to be predicted is determined based on the values of the historical power load data in each time scale; and the value band of power load data of each day in the operation mode in the next time scale to be predicted is obtained based on the dominant load curve and the confidence interval of the next time scale to be predicted.
  • Statistical techniques are mainly used in the process, and therefore it is easy to implement and can ensure prediction accuracy.
  • data of the converted three-dimensional matrix is clustered in the unit of day for long time scales, thereby making it possible to classify days with similar characteristics as one cluster, corresponding to one operation mode. Then, in each operation mode, the historical power load data in each time scale is used to predict the power load data in the next time scale, which further improves prediction accuracy.
  • the ratio of the days in each cluster to the total number of days is calculated, and clusters with the ratio smaller than a set second threshold and at the same time the operation modes corresponding to the clusters are abandoned, which reduces the calculation work for useless operation modes and saves resources for calculation and storage.
  • the historical power load data in the one-dimensional time series is converted to a three-dimensional matrix based on a longer time scale, thus providing higher applicability of the three-dimensional matrix, i.e., the complex conversion from one-dimensional data to a three-dimensional data needs to be performed only once before data in both longer and shorter time scales can be easily further converted through the three-dimensional matrix based on a long time scale, without the need to perform again the complex conversion from one-dimensional data to three-dimensional data, thereby reducing the difficulty in data processing.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Power Engineering (AREA)
  • Supply And Distribution Of Alternating Current (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
US17/614,401 2019-05-29 2019-05-29 Power load prediction method and apparatus, and storage medium Active US11740603B2 (en)

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/089102 WO2020237539A1 (zh) 2019-05-29 2019-05-29 电力负荷的预测方法、装置及存储介质

Publications (2)

Publication Number Publication Date
US20220214655A1 US20220214655A1 (en) 2022-07-07
US11740603B2 true US11740603B2 (en) 2023-08-29

Family

ID=73552455

Family Applications (1)

Application Number Title Priority Date Filing Date
US17/614,401 Active US11740603B2 (en) 2019-05-29 2019-05-29 Power load prediction method and apparatus, and storage medium

Country Status (4)

Country Link
US (1) US11740603B2 (zh)
EP (1) EP3965027A4 (zh)
CN (1) CN113853623A (zh)
WO (1) WO2020237539A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220236706A1 (en) * 2019-05-29 2022-07-28 Siemens Aktiengesellschaft Power grid user classification method and device and computer-readable storage medium

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11979023B2 (en) * 2020-01-28 2024-05-07 Enphase Energy, Inc. Method and apparatus for relative value generation of resources based on knapsack algorithms
CN114336651B (zh) * 2022-01-04 2023-08-01 国网四川省电力公司营销服务中心 一种基于削峰潜力的电力调度方法及装置
CN115423209B (zh) * 2022-09-27 2023-07-04 中国长江三峡集团有限公司 一种电力负荷需求预测方法、装置、电子设备及存储介质
CN115630831B (zh) * 2022-12-06 2023-05-12 北京华联电力工程监理有限公司 一种电力需求侧管理辅助决策支持系统
CN116257745B (zh) * 2023-05-10 2023-08-15 杭州致成电子科技有限公司 一种负荷电流极端异常数据处理方法和装置
CN116258355B (zh) * 2023-05-15 2023-08-11 国网浙江省电力有限公司永康市供电公司 适用于多点功率估计的配电台区负荷曲线分解方法及装置
CN116307944B (zh) * 2023-05-18 2023-08-04 山东欣瑞电气有限公司 基于人工智能与物联网的配电箱远程监测系统
CN116307295B (zh) * 2023-05-22 2023-08-04 南京宝能科技有限公司 应用于云平台的智慧能源数字化管理系统及方法
CN116914781B (zh) * 2023-09-12 2023-12-01 中国三峡新能源(集团)股份有限公司辽宁分公司 一种新能源快速频率响应系统及方法
CN117787652A (zh) * 2024-01-11 2024-03-29 湖北华中电力科技开发有限责任公司 园区电力需求预测方法、相关装置及存储介质
CN117951555A (zh) * 2024-03-27 2024-04-30 山东德源电力科技股份有限公司 一种用于负控安装测试装置的策略制定方法
CN117977576A (zh) * 2024-03-28 2024-05-03 国网四川省电力公司电力科学研究院 一种基于多尺度历史负荷数据的台区负荷预测方法

Citations (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030101009A1 (en) * 2001-10-30 2003-05-29 Johnson Controls Technology Company Apparatus and method for determining days of the week with similar utility consumption profiles
US6577962B1 (en) * 2000-09-28 2003-06-10 Silicon Energy, Inc. System and method for forecasting energy usage load
US20050096797A1 (en) * 2003-10-30 2005-05-05 Hitachi, Ltd. Method, system and computer program for managing energy consumption
US20120084063A1 (en) * 2009-06-22 2012-04-05 Johnson Controls Technology Company Systems and methods for detecting changes in energy usage in a building
US20120150707A1 (en) * 2010-09-20 2012-06-14 Johnson Controls Technology Company Systems and methods for providing energy efficient building equipment and services
US20120278051A1 (en) * 2011-04-29 2012-11-01 International Business Machines Corporation Anomaly detection, forecasting and root cause analysis of energy consumption for a portfolio of buildings using multi-step statistical modeling
US20120330626A1 (en) * 2011-06-24 2012-12-27 International Business Machines Corporation Estimating building thermal properties by integrating heat transfer inversion model with clustering and regression techniques for a portfolio of existing buildings
US20130127443A1 (en) * 2011-11-17 2013-05-23 Sony Corporation Electric power management apparatus and electric power management method
US20140152463A1 (en) 2012-12-03 2014-06-05 Institute For Information Industry Power monitoring device and power analysis method
CN104036356A (zh) 2014-06-12 2014-09-10 国家电网公司 一种利用分形算法对电网未来运行状态进行预测的方法
US9129231B2 (en) * 2009-04-24 2015-09-08 Rockwell Automation Technologies, Inc. Real time energy consumption analysis and reporting
US20150331023A1 (en) * 2014-05-15 2015-11-19 International Business Machines Corporation Building energy consumption forecasting procedure using ambient temperature, enthalpy, bias corrected weather forecast and outlier corrected sensor data
US20170124466A1 (en) 2015-10-30 2017-05-04 Global Design Corporation Ltd. System, Method and Cloud-Based Platform for Predicting Energy Consumption
US20170337644A1 (en) * 2016-05-23 2017-11-23 General Electric Company Data driven invocation of realtime wind market forecasting analytics
CN108074015A (zh) 2017-12-25 2018-05-25 中国电力科学研究院有限公司 一种风电功率超短期预测方法及系统
US20180255470A1 (en) * 2015-11-05 2018-09-06 Huawei Technologies Co., Ltd. Network Event Prediction Method and Apparatus and Method and Apparatus for Establishing Network-Event Prediction Model
CN109190856A (zh) 2018-10-30 2019-01-11 深圳供电局有限公司 一种区分待预测日的城市电网短期负荷预测方法及系统
CN109308541A (zh) 2018-08-02 2019-02-05 华南理工大学 一种基于场景分析的电动汽车充电负荷预测方法
CN109508830A (zh) 2018-11-15 2019-03-22 云南电网有限责任公司 一种电动汽车时空动态负荷预测的方法

Patent Citations (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6577962B1 (en) * 2000-09-28 2003-06-10 Silicon Energy, Inc. System and method for forecasting energy usage load
US20030101009A1 (en) * 2001-10-30 2003-05-29 Johnson Controls Technology Company Apparatus and method for determining days of the week with similar utility consumption profiles
US20050096797A1 (en) * 2003-10-30 2005-05-05 Hitachi, Ltd. Method, system and computer program for managing energy consumption
US9129231B2 (en) * 2009-04-24 2015-09-08 Rockwell Automation Technologies, Inc. Real time energy consumption analysis and reporting
US20120084063A1 (en) * 2009-06-22 2012-04-05 Johnson Controls Technology Company Systems and methods for detecting changes in energy usage in a building
US20120150707A1 (en) * 2010-09-20 2012-06-14 Johnson Controls Technology Company Systems and methods for providing energy efficient building equipment and services
US20120278051A1 (en) * 2011-04-29 2012-11-01 International Business Machines Corporation Anomaly detection, forecasting and root cause analysis of energy consumption for a portfolio of buildings using multi-step statistical modeling
US20120330626A1 (en) * 2011-06-24 2012-12-27 International Business Machines Corporation Estimating building thermal properties by integrating heat transfer inversion model with clustering and regression techniques for a portfolio of existing buildings
US20130127443A1 (en) * 2011-11-17 2013-05-23 Sony Corporation Electric power management apparatus and electric power management method
TW201423291A (zh) 2012-12-03 2014-06-16 Inst Information Industry 電力監控裝置及其電力分析方法
CN103853131A (zh) 2012-12-03 2014-06-11 财团法人资讯工业策进会 电力监控装置及其电力分析方法
TWI489230B (zh) 2012-12-03 2015-06-21 Inst Information Industry 電力監控裝置及其電力分析方法
US20140152463A1 (en) 2012-12-03 2014-06-05 Institute For Information Industry Power monitoring device and power analysis method
US20150331023A1 (en) * 2014-05-15 2015-11-19 International Business Machines Corporation Building energy consumption forecasting procedure using ambient temperature, enthalpy, bias corrected weather forecast and outlier corrected sensor data
CN104036356A (zh) 2014-06-12 2014-09-10 国家电网公司 一种利用分形算法对电网未来运行状态进行预测的方法
US20170124466A1 (en) 2015-10-30 2017-05-04 Global Design Corporation Ltd. System, Method and Cloud-Based Platform for Predicting Energy Consumption
US20180255470A1 (en) * 2015-11-05 2018-09-06 Huawei Technologies Co., Ltd. Network Event Prediction Method and Apparatus and Method and Apparatus for Establishing Network-Event Prediction Model
US20170337644A1 (en) * 2016-05-23 2017-11-23 General Electric Company Data driven invocation of realtime wind market forecasting analytics
CN108074015A (zh) 2017-12-25 2018-05-25 中国电力科学研究院有限公司 一种风电功率超短期预测方法及系统
CN109308541A (zh) 2018-08-02 2019-02-05 华南理工大学 一种基于场景分析的电动汽车充电负荷预测方法
CN109190856A (zh) 2018-10-30 2019-01-11 深圳供电局有限公司 一种区分待预测日的城市电网短期负荷预测方法及系统
CN109508830A (zh) 2018-11-15 2019-03-22 云南电网有限责任公司 一种电动汽车时空动态负荷预测的方法

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
FENG YONGHAN; RYAN SARAH M.: "Day-ahead hourly electricity load modeling by functional regression", APPLIED ENERGY., ELSEVIER SCIENCE PUBLISHERS., GB, vol. 170, 17 March 2016 (2016-03-17), GB , pages 455 - 465, XP029469811, ISSN: 0306-2619, DOI: 10.1016/j.apenergy.2016.02.118
Feng, Yonghan et al.; "Day-ahead hourly electricity load modeling by functional regression", Applied Energy, Elsevier Science Publishers, GB, vol. 170, Mar. 17, 2016 (Mar. 17, 2016), XP029469811, ISSN: 0306-2619, DOI: 10.1016/J.APENERGY.2016.02.118; 2016.
International Search Report and Written Opinion and English translation thereof dated Mar. 6, 2020.
Mordjaoui Mourad et al.; "Electric load forecasting by using dynamic neural network", International Journal of Hydrogen Energy, Elsevier, Amsterdam, NL, vol. 42, No. 28, Apr. 25, 2017 (Apr. 25, 2017), XP085128760, ISSN: 0360-3199, DOI: 10.1016/J.IJHYDENE.2017.03.101; 2017.
MORDJAOUI MOURAD; HADDAD SALIM; MEDOUED AMMAR; LAOUAFI ABDERREZAK: "Electric load forecasting by using dynamic neural network", INTERNATIONAL JOURNAL OF HYDROGEN ENERGY, ELSEVIER, AMSTERDAM, NL, vol. 42, no. 28, 25 April 2017 (2017-04-25), AMSTERDAM, NL, pages 17655 - 17663, XP085128760, ISSN: 0360-3199, DOI: 10.1016/j.ijhydene.2017.03.101

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220236706A1 (en) * 2019-05-29 2022-07-28 Siemens Aktiengesellschaft Power grid user classification method and device and computer-readable storage medium

Also Published As

Publication number Publication date
CN113853623A (zh) 2021-12-28
WO2020237539A1 (zh) 2020-12-03
EP3965027A4 (en) 2023-01-25
EP3965027A1 (en) 2022-03-09
US20220214655A1 (en) 2022-07-07

Similar Documents

Publication Publication Date Title
US11740603B2 (en) Power load prediction method and apparatus, and storage medium
US11831160B2 (en) Power load data prediction method and device, and storage medium
CN109495318B (zh) 一种移动通信网络流量预测方法、装置及可读存储介质
CN106685674B (zh) 网络事件预测以及建立网络事件预测模型的方法和装置
WO2019127821A1 (zh) 可视化报表的制作方法、装置、终端设备及存储介质
US20170161409A1 (en) System and Method for Modelling Time Series Data
US20220260619A1 (en) Abnormal electricity use recognition method and device, and computer-readable medium
KR102557177B1 (ko) 홀로그램 양자 동역학 시뮬레이션
US20180349801A1 (en) Log analysis system, log analysis method, and log analysis program
US11227433B2 (en) Device and method for extracting terrain boundary
US20190095495A1 (en) Computerized methods and systems for grouping data using data streams
US9788158B1 (en) Systems and methods to measure the population of mobile devices in a predefined geographical region
CN110147470B (zh) 一种跨机房数据比对系统及方法
US20230131518A1 (en) Model Generation Method and Apparatus, Object Detection Method and Apparatus, Device, and Storage Medium
US20160371408A1 (en) Population projection method and population projection apparatus
CN111680085A (zh) 数据处理任务分析方法、装置、电子设备和可读存储介质
US20180060448A1 (en) Computer and method of creating graph data
CN111163482A (zh) 数据的处理方法、设备及存储介质
US20220236706A1 (en) Power grid user classification method and device and computer-readable storage medium
US20160277494A1 (en) Synchronization control device, synchronization system, and synchronization control method
CN112287503A (zh) 用于交通需求预测的动态空间网络构建方法
CN109669919B (zh) 文件来源识别方法及装置
US10726013B2 (en) Information processing device, information processing method, and recording medium
CN111683296B (zh) 视频切分方法、装置、电子设备及存储介质
US20170337644A1 (en) Data driven invocation of realtime wind market forecasting analytics

Legal Events

Date Code Title Description
FEPP Fee payment procedure

Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

AS Assignment

Owner name: SIEMENS LTD., CHINA, CHINA

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:HUA, WEN TAO;LI, JING;LIU, HAO;AND OTHERS;SIGNING DATES FROM 20211125 TO 20211126;REEL/FRAME:058453/0604

AS Assignment

Owner name: SIEMENS AKTIENGESELLSCHAFT, GERMANY

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:SIEMENS LTD., CHINA;REEL/FRAME:058550/0403

Effective date: 20211213

STPP Information on status: patent application and granting procedure in general

Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: FINAL REJECTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: RESPONSE AFTER FINAL ACTION FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: ADVISORY ACTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION

STPP Information on status: patent application and granting procedure in general

Free format text: NON FINAL ACTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: NOTICE OF ALLOWANCE MAILED -- APPLICATION RECEIVED IN OFFICE OF PUBLICATIONS

STCF Information on status: patent grant

Free format text: PATENTED CASE